Abstract
According to Williamson (The philosophy of philosophy, Blackwell, Oxford, 2007), our knowledge of metaphysical necessities and possibilities is just a “special case” of our knowledge of counterfactual conditionals. This subsumption of modal under counterfactual thinking mainly serves a methodological role: to sign the end of “philosophical exceptionalism” in modal epistemology, namely the view that our knowledge of metaphysical modalities is obtained by means of a special, dedicated, possibly a priori, capacity. In this paper, I show that a counterfactual approach to modal epistemology is structurally similar to more traditional “conceivability-based” approaches. On this basis, I then show that the counterfactual approach suffers some of the same problems and I conclude that it is still based on a quite exceptional capacity to determine the truth of modal metaphysical claims. Given that, for Williamson, the epistemology of thought experiments should also be subsumed under the counterfactual approach, the problem I raise in this paper has consequences for his approach to thought experiments.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
I am here assuming, as Williamson does, that to think counterfactually is to think some counterfactuals and thus that the development of an epistemology of counterfactual thinking is just the development of an epistemology of counterfactuals.
One could also claim that the counterfactual approach to modal epistemology may help us to “reform” the traditional, aprioristic conception of conceivability, rather than to dispense with it. This would surely be a plausible move, I am myself sympathetic with it, and Williamson himself seems to give some indications in this direction (e.g., Williamson 2007, p. 164). As far as this paper is concerned, however, I will use “conceivability” as the name of the traditional, aprioristic, exceptional faculty postulated by rationalists to investigate the modal realm, “conceivability-based view” as the name of the modal epistemology based on such an exceptional faculty and I will assume that the counterfactual view of modal epistemology is an alternative to it.
See Williamson (2007, p. 206).
Note that the acceptance of (Count-Nec) depends on the acceptance of a certain view about the vacuous truth of counterfactuals. In particular, one has to accept the idea that counterfactuals conditionals with impossible antecedents are vacuously true: if ϕ is necessary, \(\lnot \phi\) is impossible and it is only by assuming that any counterfactual with an impossible antecedent is true that we can show that
is true.Kroedel (2012, p. 5).
Williamson (2007, p. 136).
Williamson seems to endorse this view and uses the term “byproduct”; (cf. 2007, p. 162):
the capacity to handle metaphysical modality is an “accidental” byproduct of the cognitive mechanisms that provide our capacity to handle counterfactual conditionals.
A referee invites me to specify better the relations between Byproduct on the one side and Performance and Reconstruction on the other. Byproduct seems to be compatible both with Performance and with Reconstruction. However, a situation where we explain our understanding of modal metaphysical claims through a combination of Performance and Byproduct would still be a situation where we have problems to explain why we seem more reliable in our understanding of some modal claims than in our understanding of their corresponding counterfactual claims. On the contrary, a combination of Reconstruction and Byproduct could be, as I said in the text, a way to give an evolutionary justification to our reconstructive hypotheses. Surely, Reconstructionalone is not enough as an explantion. So, Performance + Byproduct seems to be a possible but not an advisable combination, Reconstruction + Byproduct seems to be possible and, as far as one is interested in a reconstructive hypothesis, an advisable combination.
A suppositional approach to counterfactuals (and to conditionals in general) is defended, for example, by Dorothy Edgington (1995). According to such a view, counterfactuals are, roughly, acts of stating relative to a supposition. A counterfactual judgement involves two propositions that never combine into a single proposition: one is the proposition expressing the content of the supposition, the other is the proposition expressing the content of the judgment made under that supposition.
Notice thus that, especially within Lewis’s semantics, the definability of \(\Box\) in terms of
should simply be seen as a case of a definition of a necessity operator in terms of another one. In this sense, the definition of modal metaphysical operators in terms of counterfactual conditionals should be seen as being more similar to the definition (or a reduction) of a relative necessity in terms of another, rather than a definition of metaphysical necessity in other terms.Lewis, the most authoritative defender of the categorical interpretation of counterfactuals, avails himself of a suppositional talk, for example when he presents the antecedent of a counterfactual as a “counterfactual supposition” or when, discussing counterfactuals with impossible antecedents he qualifies the antecedents of such counterfactuals as “non-entertainable suppositions”. Cf. for example, Lewis (1973, pp. 16, 27) or also Lewis (1979, pp. 455, 462).
Maybe, on the basis of this, one could try to defend the (quite radical) view that neither modal metaphysical claims are truth-conditional.
This distinction is to be found in Barnett (2009), fn. 3. Barnett claims that both Lewis (1973) and Stalnaker (1968) could be seen as implicitly endorsing this distinction. He is actually a defender of a suppositional approach, so he denies that this distinction is tenable in the end. He thinks that the suppositional and the categorical approaches to counterfactuals make incompatible predictions and so, even postulating the distinction between interpretation and evaluation, we would have cases where we evaluate as true some counterfactuals that we would interpret as false (or viceversa).
I have trying to replicate, for the notion of “counterposing”, the case of “schmidentity” developed by Kripke (1980, p. 108).
I have chosen the example of Kennedy’s assassination because this discussion is somewhat reminiscent of the discussion about the putative distinction between counterfactual and indicative conditionals. In that context the possible difference in truth-value between (i) “If someone did not killed Kennedy, someone else did” and (ii) “If Oswald had not been killed Kennedy, someone else would have” (where all those knowing about Kennedy’s assassination would agree with the former, but some of them would not agree with the latter) is taken as evidence to show that there are two different kinds of conditionals Lewis (1973, p. 3). I think, however, that even in this case, as D. Edgington claims, the argument from the difference in truth-value to the postulation of two sorts of conditionals is “resistible”. For her, the difference marked by (i) and (ii) above is:
...maybe more like the difference between mature cheddar and freshly-made cheddar than the difference between chalk and cheese. As time passes but relevant information stays the same, “If he eats the apple, ...”, “If he were to eat it, ...”, “If he ate it ...” and “If he had eaten it, ...” may all express the same conditional thought. But the passing of time may bring new relevant information. (Edgington 2004, p. 239)
I am definitely on Edgington’s side on this issue (see pages 237–247 of her 2004’s paper for a structured defence of the view). Difference in truth-condition between an indicative and the corresponding counterfactual conditional is not enough to motivate a semantic ambiguity in “if”. I think that the case of “supposing” is even more clearly a case where postulating a semantic ambiguity is not a plausible answer.
See Williamson (2007, p. 145–155).
For the sake of this paper, I am abstracting away from some differences one could be willing to do between “supposing”, “conceiving” and “imagining”. See, for example, Szabò Gendler (2000).
Here is a more explicit version of the argument for the case of metaphysical necessity:
(2) is the claim that, if a contradiction is derivable from the counterfactual supposition that \(\lnot \phi\), then the truth of
is known. In general, it is true that, for every agent i, if i is able to determine that from the counterfactual supposition that \(\lnot \phi\) it follows a contradiction, then i knows the truth of . The role of (2) is thus simply that of connecting the capacity to determine the truth of a counterfactual with knowledge of such a counterfactual. The case of metaphysical possibility is structurally analogous.Here some quotations from (Chalmers 2002, p. 149):
The central sort of negative conceivability holds that p is negatively conceivable when p is not ruled out a priori, or when there is no (apparent) contradiction in p.
We can say that p is ideally negatively conceivable when it is not a priori that \(\lnot p\).
Of course, there are counterfactuals whose truth is determined a priori such as “If twelve people had come to the party, more than eleven people would have come to the party” (the example is taken from Williamson 2007, p. 143), but this is surely not a typical case of a counterfactual. Furthermore, I am not really sure that the way in which this kind of counterfactuals are evaluated is exclusively a priori: after all, even mathematical facts have metaphysical consequences.
Of course, this is true just in case the difference is accepted between what is a priori possible or necessary and what is metaphysically possible or necessary. Since Kripke (1980) everyone tends to accept this distinction. If the further assumption is made that what is a priori possible is weaker than what is metaphysically possible (namely that there are possibilities a priori that are not metaphysically possible), then, by negative conceivability, we could end up as counting as possible some proposition that is metaphysically impossible.
Notice, furthermore, that many of those who traditionally embraced something like the negative conceivability view were interested to access just some form of a priori knowable, logical or conceptual, possibility and necessity. Conceivability tests are within these approaches, just tests about conceptual consistencies. In this paper, I am discussing just the position of those who claim that, by means of negative conceivability, we can have access to the metaphysical modalities.
For some, an argument is valid if and only if there is an “a priori route” from premises to conclusion. Cf. Edgington (2004, p. 10). However, to make my point it is only sufficient to assume that every case of logical entailment is a case of a priori entailment.
Note that there can be non-exceptional capacities based on the a priori that may be associated with a realist conception of the phenomenon tracked by the capacity: for example, one can hold that our knowledge of mathematics is a priori, but that not all mathematical truths can be known, that there are elusive mathematical truths. Being known a priori is thus not the exclusive mark of the exceptional.
On the importance of the notion of metaphysical entailment in the evaluation of counterfactuals (although in a context of a suppositional approach to counterfactuals), see Barnett (2011).
As Stalnaker puts it: “[in Lewis’s semantics], the antecedent of a counterfactual acts like a necessity operator on its consequent” (Stalnaker 1978, p. 93).
The reasoning, assuming that there is an agent with the relevant capacity, is the following:
(2) is, again, the thesis that, if an agent is able to determine that the counterfactual supposition that ϕ metaphysically implies a contradiction, then such an agent knows the truth of
. As I said in the text, I think that (2) is plausible even without assuming the existence of idealized agents and it seems to be true for every agent.The derivation is the following:
Notice that the modalities involved in this derivation are not necessarily metaphysical ones, because only a system as strong as S4 is needed; thus, the relevant possibility could be less strong than metaphysical possibility. Of course, with weaker forms of necessity than metaphysical necessity, the possible existence of a becomes less appealing.
Notice, however, that to deny the possibility of a is to claim that it is metaphysically impossible that there is an agent that, under the assumption that ϕ is metaphysically necessary, is able to determine that the counterfactual supposition that \(\lnot \phi\) metaphysically entails a contradiction, for every ϕ. At least prima facie, I do not find anything contradictory in the possible existence of an agent like a, nor I can imagine an argument aiming to deny a’s metaphysical possibility.
(the symbol “
.
should simply be seen as a case of a definition of a necessity operator in terms of another one. In this sense, the definition of modal metaphysical operators in terms of counterfactual conditionals should be seen as being more similar to the definition (or a reduction) of a relative necessity in terms of another, rather than a definition of metaphysical necessity in other terms.
References
Barnett D (2009) The myth of the categorical counterfactual. Philos Stud 144:281–296
Barnett D (2011) Counterfactual entailment. Proceedings of the Aristotelian society, CXII, pp 73–97
Chalmers D (2002) Does conceivability entails possibility? In: Gendler J, Hawthorne TS (eds), Conceivability and possibility, 22 Oxford University Press, Oxford, pp 145–200
Cohnitz D (2012) The logic(s) of modal knowledge. In: Restall G, Russell G (eds), New waves in philosophical logic, Palgrave Macmillan UK, London, pp 63–83
Edgington D (1995) On conditionals. Mind 104:235–330
Edgington D (2004) Two kinds of possibility. Aristot Soc Suppl 78:1–22
Gendler TS (2000) Thought experiment: on the powers and limits of imaginary cases. Routledge, New York
Gendler TS (2004) Thought experiments rethought — and reperceived. Philos Sci 71(5):1152–1163
Kripke SA (1980) Naming and necessity. Harvard University Press, Harvard
Kroedel T (2012) Counterfactuals and the epistemology of modality. Philos Impr 12(1):1–14
Lewis DK (1971) Completeness and decidability of three logics of counterfactual conditionals. Theoria 37:74–85
Lewis DK (1973) Counterfactuals. Blackwell, Oxford
Lewis DK (1979) Counterfactual dependence and time’s arrow. \(No\hat{u}s,\) 13, 455–476
Menzies P (1998) Possibility and conceivability: a response-dependent account of their connections. Eur Rev Philos 3:255–277
Norton JD (1996) Are thought experiments just what you thought. Can J Philos 26(3):333–366
Norton JD (2004) Why thoughts experiment do not transcend empiricism. In: Hitchock C (ed), Contemporary debates in philosophy of science, Blackwell, Oxford, pp 44–66
Stalnaker R (1968) A theory of conditionals. In: Rescher N (ed) Studies in logical theory, Blackwell, Oxford, pp 98–112
Stalnaker R (1978) A defense of conditional excluded middle. In: Harper W, Stalnaker R, and Pearce G (eds) Reidel, Dordrecht, pp 87–103
Szabò Gendler G (2000) The puzzle of imaginative resistance. J Philos 97(2):55–81
Williamson T (2007) The philosophy of philosophy. Blackwell, Oxford
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Morato, V. Conceivability, Counterfactual Thinking and Philosophical Exceptionality of Modal Knowledge. Topoi 38, 821–833 (2019). https://doi.org/10.1007/s11245-017-9464-x
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11245-017-9464-x